Home > Why Do Your Ads Keep Finding the Wrong People?

Why Do Your Ads Keep Finding the Wrong People?

Man in glasses filling out an online contact form on a laptop, with a cup of coffee on the table beside him

By Adrienne Wilkerson | Published September 28, 2026 | Last updated September 28, 2026

Your ads are learning exactly what you taught them. That’s the problem.

Ad platforms like Google and Meta optimize toward whatever event you tell them counts as a conversion. If every form fill counts, including the person whose insurance you don’t take, the platform gets better at finding more people who fill out forms. If only a right-fit inquiry counts, it gets better at finding right-fit people. The budget is the same either way. What you’re buying is completely different.

I’ve been thinking about this all month, because it’s where every conversation we’ve had in September ends up. We talked about marketing and intake disagreeing about what a good lead even is. We talked about which marketing actually filled your schedule. We talked about the first phone call. All of that work happens inside your practice. This week is about what happens next: getting what your team learns about each lead back to the ad platforms you’re paying every month. Your marketing partner should be the one handling the platform side. Your part is making sure what intake learns actually reaches them.

Because in most practices, it never does. The ad platforms get funded every month, and the one piece of information that would make them smarter stays in the CRM.

What are your ads actually learning from?

Your ads learn from the conversion action you mark as primary, and in most behavioral health accounts that action is a form submission or a phone call. Google’s own documentation says primary conversion actions are the ones used for bidding. Everything the algorithm does with your money is aimed at producing more of that one event.

Here’s the thing. A form fill tells the platform that somebody filled out a form. That’s all it knows.

It doesn’t know that intake called them back and found out they have a plan you’ve never been in-network with. It doesn’t know they were looking for a service you don’t offer. It doesn’t know they booked, didn’t show, and never answered again. From the platform’s point of view, that person was a win. So it goes looking for more people who look like them.

Google says as much in its guidance on primary and secondary conversion actions: primary actions feed bidding, secondary actions are there for you to watch. Most of the practice owners I talk to have never opened that setting. Honestly, most of them don’t know it exists. You went to school to help people. Conversion goals were never on the syllabus.

But you’re still the one signing off on the spend.

“Your ad platform is obedient. It will spend every dollar you give it chasing whatever you told it a win looks like.”

Why do the wrong people keep filling out your form?

The wrong people fill out your form because they’re hunting, and in behavioral health the hunt is brutal. Someone who needs care often has to contact several practices before one takes their insurance, and every one of those attempts looks like a conversion to the ad platform that sent them.

We see this with our clients constantly. A practice is clear on its website about which plans it accepts. It’s on the insurance page. It’s in the FAQ. Sometimes it’s on the contact page itself. And the form fills still come in from people with plans the practice doesn’t take.

I get frustrated about it on the client’s behalf, and then I remember what the person on the other side is up against. They’re exhausted, and they’re filling out every form they can find.

The data backs that up. In an October 2025 data brief, the HHS Office of Inspector General found that Medicare Advantage plans’ networks included, on average, just 16 percent of a county’s behavioral health workforce, and Medicaid managed care plans averaged 31 percent. The same report found that on average, 55 percent of the behavioral health providers listed in Medicare Advantage networks didn’t provide a single service to enrollees in 2023.

Think about what that means for the person searching. The directory their plan gave them is full of names that won’t work. So they go to Google. They click, they fill out a form, they move on to the next one, and they do it again.

The commercial side isn’t much better. An April 2024 RTI International study found that patients went out-of-network 3.5 times more often for a behavioral health clinician than for a medical or surgical one, and 10.6 times more often for a psychologist. The study’s lead author, Tami L. Mark, Ph.D., summed up the network gap this way: health plans “are not using the same measures” to strengthen their behavioral health networks that they use for medical ones.

So when a person with the wrong insurance lands on your form, blame the system they’re stuck in. Your part of the problem starts when your ad account counts them as a success and pays to find more of them.

What happened when we changed what counted as a win?

When we changed what counted as a conversion, the platforms started getting a cleaner signal, because the only people reaching the finish line were people the practice could actually serve. The fix started as a form fix. It turned into an ad fix without anyone having to rebuild the campaign.

Here’s what we do for clients who keep getting inquiries they can’t take.

We add one question near the top of the contact form: which insurance do you have? Usually it’s a drop-down with the plans the practice accepts, plus “Other” and “I don’t have insurance.”

If the person picks one of the listed plans, the form keeps going like normal.

If they pick “Other” or “none,” they land on a page that tells them, kindly and plainly, that the practice doesn’t accept their insurance. Then it gives them a choice. They can stop right there, or they can keep going as a self-pay client.

And here’s the part that matters for your ad budget. The conversion only fires when someone finishes on a qualified path. The in-network person who completes the form counts. The self-pay person who chooses to keep going counts. The person who learns it’s not a fit and stops there doesn’t count, so the platform never learns to go find more of them.

Nobody got turned away rudely. Nobody wasted twenty minutes waiting for a callback that was always going to be a no. And the ad platform quietly started learning from a much more honest signal.

We test things on Beacon first before we roll them out to clients, and this was no different. Beacon has a threshold for the practices we can serve well, based on budget and team size. It’s on our website. We’re explicit about it. And we still get inquiries from practices that are well under it.

So our own form works the same way now. If someone selects a budget or staff size below our threshold, they go to a separate form that tells them we’re happy to talk, but they’re under our threshold and we may not be the best fit. They get to decide whether to keep going.

I will be the first to admit that felt a little uncomfortable at first. Nobody in business loves writing a sentence that tells a potential client “maybe not us.” But people hate being sold. They love being told the truth. And letting someone self-select out is one of the most respectful things a form can do.

“The kindest thing your contact form can do is tell someone the truth before they spend twenty minutes waiting on a callback that was always going to be a no.”

Is a smarter form enough, or do you need to send lead quality back to the platform?

A smarter form is the right first step, but it only filters what someone can know about themselves at the moment they fill it out. Insurance is knowable at the form. Whether a person is a good clinical fit, whether they show up, and whether they stay are not. Those answers live in your CRM, and getting them back to the ad platform is a separate job.

That’s where offline conversions come in. Your intake team grades the lead in the CRM, and an approved signal goes back to the platform that says, in effect, “this one was worth having.” Stephanie wrote the mechanics of that this week in how to send lead quality back to your ad platforms, including what should never leave your building. In behavioral health, that last part isn’t optional. Your CRM knows things about people that an ad platform has no business holding.

What I want owners to hold onto is the order of operations.

  • Start with what’s knowable at the form. Insurance, location, service type, self-pay willingness. If a question would let someone self-select out, ask it before the conversion fires.
  • Grade leads the same way every time. You can’t send quality back if nobody recorded it. This is the whole reason we spent September on definitions and CRM fields.
  • Pick a milestone the platform can actually learn from. This is where owners get tripped up. Google recommends measuring Target CPA performance over the last 30 days, with at least 30 conversions. If you only count “became a client,” a practice with 10 or 15 staff may never hit that in a month. A qualified lead or a scheduled first session is often the more useful signal.

That third one is a judgment call, and it’s different for every practice. Go too broad and you’re back to rewarding form fills. Go too narrow and the platform doesn’t have enough to learn from. Somebody has to own that call and revisit it.

“Every month you don’t send lead quality back, you’re paying full price for an algorithm that’s only allowed to learn half the lesson.”

What should a practice owner actually ask for?

A practice owner should ask what the ad account is optimizing toward, and whether that event means a right-fit person or just a completed form. That one question tells you more about whether your ad spend is working than a stack of monthly reports.

You don’t need to learn the platforms. You need to ask better questions of the people running them. Here are the ones I’d bring to your next marketing meeting.

  • “What is our primary conversion action, exactly?” If the answer is “all form submissions,” that’s your starting point.
  • “Can someone who isn’t a fit still trigger a conversion?” If yes, ask what it would take to stop that.
  • “What happens to our lead grades after intake?” If they stay in the CRM and never reach the ad account, the ads are only learning from the website.
  • “What does it actually cost us to get a client?” That number is almost always higher than cost per lead, and it’s the only one that tells you what the budget bought.

If you want to go deeper, we’ve answered a few of these on their own: what marketing metrics should a practice owner review monthly, what is cost per acquired client in behavioral health, and why is cost per lead a misleading metric. We’ve also written about why lead volume doesn’t predict revenue, which is the same problem seen from the reporting side.

One warning. When you tighten what counts as a conversion, the conversion number in your account goes down. It’s supposed to. You stopped counting people who were never going to become clients. If whoever runs your ads panics at that dip, or you do, you’ll undo the fix before it has a chance to work. Decide ahead of time that you’re going to judge this on qualified leads and new clients, and give it room.

That’s the kind of setup our paid advertising team builds and watches for clients every day, because the details of getting it right, and keeping it compliant, add up fast.

Why does this matter beyond the budget?

This matters beyond the budget because every conversion you count is a real person who reached out on what might have been a very hard day. When your system rewards the wrong reach-outs, it spends your money sending more people toward a door that was never going to open for them.

I think about that person a lot. They finally worked up the nerve to look for help. They found a practice that looked right. They filled out the form, and then they waited. Maybe a day. Maybe three. And the call they got back was, “I’m sorry, we don’t take your plan.”

Nobody did anything wrong in that story, and it still hurts. Multiply it by every ad dollar pointed at the wrong audience and you’ve got a practice that’s spending money to create disappointment. That’s not what anyone got into this field to do.

When you fix what counts, two things get better at once. Your budget stops funding misses, and the people who aren’t a fit find out sooner, from you, kindly, while they still have the energy to keep looking. The better the connection, the more likely somebody is to become a client, or to recognize you’re not the right fit and move on without losing a week.

“Good behavioral health marketing helps the wrong-fit person find their next door faster.”

That’s what I mean when I say marketing is human-to-human connection. The conversions follow when you get the connection right. And the platforms you’re paying every month can only learn what a real connection looks like if someone tells them.

So here’s what I’d love to know from you. What’s the one question your intake team asks on every first call that you wish your ads already knew the answer to?

Frequently asked questions

Why do my ads bring in leads with insurance we don’t accept? Ad platforms optimize toward whatever counts as a conversion in your account. If every form submission counts, the platform learns to find more people who submit forms, including people with plans you don’t take. Adding an insurance question before the conversion fires, and only counting qualified completions, gives the platform a cleaner signal to learn from.

Should a practice stop counting all form fills as conversions? Most practices should narrow what counts, but not all at once. Start by making sure someone who clearly isn’t a fit, such as an out-of-network inquiry who chooses not to continue, can’t trigger the conversion. Then decide whether qualified leads or scheduled sessions give the platform enough volume to learn from.

How many conversions does an ad platform need to learn? It depends on the platform and bid strategy. For Target CPA, Google recommends evaluating performance over the last 30 days with at least 30 conversions. Smaller practices often need a milestone earlier than “became a client,” such as a qualified lead, so the platform has enough signal to work with.

Is it safe to send lead quality back to Google or Meta in behavioral health? It can be, with care. The platform should receive only a minimal, approved conversion signal, never diagnosis, treatment, insurance, or clinical details. Event names should be generic. Because protected health information may be involved, your compliance or legal team should approve the workflow before anything goes live.

Will my conversion numbers drop if I change what counts? Yes, and that’s expected. When you stop counting inquiries that were never going to become clients, the conversion total falls. Judge the change on qualified leads, new clients, and cost per acquired client over a reasonable window instead of the raw conversion count.


This article is general information about marketing and advertising practices and is not legal or compliance advice.

Prefer to Listen? Tune into The Beacon Way Podcast

On The Beacon Way, Adrienne Wilkerson, CEO and co-founder of Beacon, connects with entrepreneurs and business leaders who share what it really takes to build and lead.